CA3034865A1 - Photonic device and conversion process for double band frequency - Google Patents

Abstract

A photonic frequency converter comprising: an electro-optical intensity modulator (MEO) having an optical input (EO), an optical output (SO) and at least one RF input (EC, EC1, EC2) for receiving two radio frequency signals of modulation (V RF1, V RF2) at different frequencies; a set of optical sources configured to generate optical signals (SO1 ¨ SO4) at at least two different wavelengths (.lambda.1 ¨ .lambda.4), modulated by local oscillator signals (SOL1 ¨ SOL4) respective of which at least two have different frequencies; and an optical multiplexer arranged to multiplex the said optical signals and inject them into the optical input of the modulator. Frequency conversion method using such a converter

Description

PHOTONIC DEVICE AND METHOD FOR CONVERTING
DOUBLE BAND FREQUENCY
The invention relates to a frequency converter dual-band photonics, as well as on a process and a payload frequency conversion photonics using such a converter.
The invention relates to the field of photonics. She is mainly, but not exclusively, intended to be applied to the space telecommunications. In particular, it aims to reduce the complexity, weight, bulk, cost and / or consumption telecommunications payloads embedded in satellites, including receivers used at the end of links rising. However, this concept is also applicable for the conversion to a higher frequency for the transmission of signals to radio frequency (RF), especially in downlinks.
In the field of telecommunications it is often necessary to perform signal frequency conversions to radiofrequency, and in particular microwave. In the field of space telecommunications, frequency conversions are, for example, example, necessary to retransmit on a downlink signals received via an uplink. This can also be used to lower the frequency of a received signal to facilitate its processing by electronic circuits, or conversely to increase the frequency of a signal for transmission. Finally, it may be a question of combining radiofrequency signals from different sources whose spectra are overlap, in which case a frequency conversion is required to avoid interference. In a conventional way, each operation of frequency conversion requires the use of a mixer to separate radio frequency, powered by a local oscillator. When the number signals to be processed is important, this can lead to the realization of very complex, and therefore expensive, heavy frequency conversion systems,

2 bulky and power consumers - these last three parameters being particularly penalizing in space applications.
By way of example, FIG. 1 schematically illustrates the structure and operation of a frequency conversion device to microwaves according to the prior art. This device receives as input four radiofrequency signals, for example two (V ,, V2) in V-band (40 - 75 GHz) and two others (Ka1, Ka2) in Ka band (26.5 - 40 GHz). After having been amplified and filtered by respective amplification and filtering circuits, AF1, AF2, AF3, AF4, these signals are provided on first inputs of four respective microwave mixers, MX1, MX2, MX3, MX4. Each of these mixers also receives, on a second input, a mono-frequency generated by a respective local oscillator LOivk, LO2vk, LO3vk, LOavk.
At its output, each mixer provides a signal obtained by mixing - that is ie by multiplying between them - the signals present at these inputs.
Concretely, we recover at the output of each mixer a signal composite containing a first component corresponding to the signal present on the first entry but whose spectrum is shifted upwards frequency equal to that of the local oscillator signal, and a second component also corresponding to the signal present on the first input but whose spectrum is shifted down)> frequency of local oscillator. Typically, a filter (not shown) allows select only one of these components. In the example of Figure 1, the four selected output signals (Kaid, Ka2d, Ka3d, Kaki) belong to the Ka band. The device therefore operates two conversions of the band V to the Ka band and two conversions within the Ka band.
A disadvantage of the device of Figure 1 is that, for operate four conversions, it must necessarily use four mixers RF separate. Indeed, if two signals belonging to the same band (VI and V2, or Kai and Ka2) were injected into the same mixer, one would recover in output of spurious signals corresponding to intermodulation products.
In addition, the useful bandwidth of a microwave mixer is quite narrow and does not allow to operate on signals belonging to bands

3 of different frequency. The payloads of the satellites of telecommunications are generally designed to perform several dozens of frequency conversions, and therefore require a number corresponding mixers. This has a very significant impact on critical parameters of the payload: its mass, its bulk, and its energy consumption.
It is known to transpose several radiofrequency signals on respective optical carriers and then convert them to frequency in using electro-optical modulators in place of radio frequency system of Figure 1. When using a modulator for each signal to be converted, this approach - which can be described as photonics)> ¨ provides only a limited reduction in the mass, congestion and / or energy consumption of the payload. She allows in some cases to obtain in certain cases a greater purity spectral output conversion, and to be able to carry the signal by fiber optical, but it requires as many mixers as there are RF signals to treat.
In some cases, however, a photonic approach allows a reduction in the number of modulators, ie mixers. By For example, it is known to transfer radiofrequency signals onto optical carriers of different wavelengths, to multiplex them and to apply them to the input of a single modulator. See for example the document FR 3,043,514. This approach, however, can only be used when it is desired to apply the same frequency shift to several RF signals. It is of no use when several signals to radiofrequency must undergo different frequency offsets.
Another photonic approach is to transfer a plurality of single-frequency signals of local oscillators on carriers at different wavelengths, to multiplex them, and to apply to the input of an electro-optical modulator to mix them at a radio frequency signal. See for example document FR 2 864 385. This approach is appropriate only in the case of simultaneous application

4 several frequency offsets to the same radio frequency signal. She is also of no use when several radio frequency signals must have different frequency offsets.
The invention aims to overcome, in whole or in part, the aforementioned drawbacks and limitations of the prior art. More particularly, it aims to provide a device for achieving, on a plurality of distinct RF signals, frequency offsets that may be different, and mass, energy consumption and / or reduced size compared to the prior art.
According to the invention, this object is achieved by resorting to a photonic technique, in which several single-frequency signals local oscillators are transferred on optical carriers to different wavelengths, multiplexed on the same optical fiber, then mixed with two RF signals using the same electronic modulator.
optical. This makes it possible to divide the number of mixers by at least one factor two, which makes such a photonic solution even more advantageous. Another advantage of the invention, or at least some of its embodiments, is its generic character: indeed, one and the same Frequency converter can be used in several spectral bands RF. Moreover, the use of a photonic solution according to the invention allows to obtain, under particular polarization conditions of the modulator, better spectral purity in terms of frequency lines in the case of a purely radiofrequency solution: more particularly, parasitic intermodulation products can be rejected more effectively.
An object of the invention is therefore a frequency converter photonics including:
an electro-optical modulator intensity having an optical input and at least one optical output;
a set of optical sources configured for generate optical signals at at least two different wavelengths,

5 modulated by respective local oscillator signals of which at least two have different frequencies;
an optical multiplexer arranged for mull: iplexer said optical signals and injecting them into the optical input of the modulator;
an optical / electrical detection system configured for convert optical signals from the electro-optical intensity modulator towards the radiofrequency domain; and an electric filter bank configured to extract spectral components of signals converted to the domain 113 radio frequency;
characterized in that the electro-optical intensity modulator also has at least one input for a DC voltage and least two RF inputs to receive two radiofrequency signals from modulation at different frequencies.
According to different embodiments of such a converter:
Said electro-optical intensity modulator may be interferometer type of Mach-Zehnder, and more particularly:
- double Mach-Zehnder interferometer type RF access and has two RF inputs for two radiofrequency signals from respective modulation; or - Mach-Zehnder I / O interferometer type and presents two RF inputs for two radio frequency signals of respective modulation; or of Mach-Zehnder interferometer type with access RF single presenting a single RF input, the converter also comprising a configured RF signal combiner to receive as input two modulation RF signals respective, combine and apply them to said RF input of electro-optical modulator of intensity.
- The converter can also include a circuit of polarization configured to apply to at least one control input of the electro-optical modulator of intensity a suitable bias voltage

6 to minimize the intensity of a spectral component of the optical signals from the electro-optical intensity modulator at a sum frequency or difference in the frequencies of the modulation radiofrequency signals.
- The converter may also include a optical demultiplexer arranged to demultiplex the optical signals from of the electro-optical intensity modulator.
Said electro-optical intensity modulator may be complementary dual optical output type and conversion system optical / electrical may comprise at least one balanced photodetector.
Another object of the invention is a payload of telecommunication for a satellite comprising at least one such converter frequency.
Yet another object of the invention is a method of frequency conversion by means of such a converter, including the following steps :
- apply to the RF input or inputs of the modulator electro-optical intensity of the converter two radiofrequency signals of which the frequencies must be converted;
- activate the optical sources of the converter and generating said optical signals at least two wavelengths different, modulated by said respective local oscillator signals - recover optical signals from the optical output of the electro-optical intensity modulator, convert them to the domain radiofrequency and filter them to extract spectral components each corresponding to a mixed input radio frequency signal with one of said local oscillator signals.
The method may also include the following step:
- apply to at least one control input of the electro-optical modulator of intensity a suitable bias voltage to minimize the intensity of a spectral component of RF signals from of the optical / electrical detection system at a sum frequency or difference in the frequencies of the modulation radiofrequency signals.

7 Other features, details and advantages of the invention will appear on reading the description made with reference to the drawings annexed given by way of example and which represent, respectively:
FIGS. 2A to 2D, diagrams of different electro-optical intensity modulators that may be suitable for use of the invention;
- Figure 2E, a graph illustrating the function of modulation transfer of an electro-optical modulator of intensity to be suitable for the implementation of the invention;
- The Figures 3, 4, 5 and 6, diagrams of converters of photonic frequency according to four embodiments of the invention;
FIGS. 7A, 7B, 7C, 8 and 9, graphs illustrating the performance of photonic frequency converters according to different embodiments of the invention;
- Figures 10 and 11, the application of a plurality of photonic frequency converters at a satellite payload of telecommunications; and - The Figure 12, the diagram of a converter according to a mode particular embodiment of the invention, simultaneously realizing a conversion up (up-conversion) and down (down-conversion).
In this document, the term radiofrequency or RF will be understood in a broad sense, covering all the frequencies in the range 3 kHz ¨ 300 GHz. Radio frequencies include microwaves, or microwaves, which correspond to the frequencies between 1 GHz and 300 GHz.
The invention will be described with reference to modes of realization using some types of electro-optical modulators intensity (in the following we will speak simply of electro-optical modulator in omitting the intensity accuracy) to Mach-Zehnder interferometer, but more generally any type of electro-optical modulator having a

8 number of accesses greater than one may be suitable for the implementation of the invention. Before describing different modes of implementation of the invention, it is appropriate to recall the operating principle of different types of Mach-Zehnder interferometer modulators that can be used for this purpose implemented.
Figure 2A illustrates a modulator MEODA (Modulator Electro-Optic Dual Access RF (dual-drive in English).
The light, polarized rectilinearly, is injected at the entrance optical component. It is propagated in a planar waveguide realized in a crystalline material E0, such as lithium niobate (LiNbO 3), with a refractive index that varies linearly with the electric field applied (Pockels effect), a Y coupler divides this waveguide in two parallel arms between which the luminous flux is distributed in a balanced. Two pairs of electrodes are arranged on both sides of each of the arms of the interferometer; a first pair of electrodes is connected to a first input, or RF port EC1 and a second first pair of electrodes is connected to a second RF input EC2. A
voltage, constant (DC, of the English direct current, that is current continuous) and / or variable, applied to a control entry modifies the index of refraction of the material, and thus the phase of the light that runs through the arm corresponding causing a variation of the optical path. Both arms are recombined by a second Y coupler, which allows the light traveling through them to interfere (the modulator is therefore essentially a interferometer). Thus, the optical intensity taken at the SO output of the modulator depends on the phase difference between the two arms of the device, and therefore on the voltages applied to the corresponding electrodes.
Generally, RF dual access modulators different RF and DC electrodes. The most widespread structures present either a DC electrode on each of the arms of the interferometer if although two separate DC voltages can be applied distinctly on each arm of the interferometer, a single electrode

9 DC on two arms of the interferometer. In this case, a single DC voltage is applied to the component. This is the most general case.
In the first embodiment with a modulator similar to that of Figure 2A, a first bias voltage keep on going VB1 generated by a polarization circuit POLAR and a first signal to radiofrequency VRF1 are applied respectively to the first input of DC control and at the first RF input; and a second DC voltage of VB2 polarization generated by the polarization circuit POLAR and a second RF signal VRF2 are respectively applied to the second control input DC and the second RF input.
The polarization voltages delivered by the circuit of POLAR polarization make it possible to set the operating point of the modulator. Figure 2D shows graphs of the transfer function of FTMI intensity modulation and modulation transfer function in FTME electric field depending on the voltage V = IVB1-VB2I. We notice that these modulation transfer functions are sinusoidal. The voltage V7, corresponds to a phase shift of lu radians between the two arms of the interferometer. The field transfer function is linear around a bias voltage Vic (more generally kV7 ,, with k integer), while than the intensity transfer function is linear around a voltage of polarization Vid2 = 1/2 V (more generally, (k + 1/2) Vn, with k integer). The choice the operating point, and therefore the bias voltage, is very important; as will be illustrated in FIGS. 7A ¨C and 8. Indeed, this choice significantly affects the performance of the converter frequency according to the invention. For this reason, the polarization circuit POLAR
generating DC voltages, preferably includes a closed-loop control to control drifts, including thermal or related to the aging of the component.
In the case of Figure 2A it was considered that two DC bias voltages are applied to control inputs lo respectively. Generally, for the components proposed on the market a single polarization voltage (VB in FIG. 2C) is applied to the modulator via a single DC control input.
Figure 2B shows the block diagram of a modulator MEOAU with unique RF access (single-drive in English). This modulator differentiates from the previous one in that it comprises a single RF input, to apply the same RF signal ¨ with reversed polarity ¨ and a bias voltage VB to the two pairs of electrodes associated with two arms of the interferometer. This signal is obtained by combining the two RF signals, VRF1 and VRF2, whose frequencies are to be converted, at means of an RF combiner noted CRF. With reference to FIGS. 8 and 9, the use of such a device for the implementation of the invention does not allows not to achieve optimal performance. In addition, the need to use RF combiners limit the bandwidth of the device and remove any Frequency genericity at this type of frequency converter Alternatively, a single pair of electrodes may be provided to apply the RF signal and the bias voltage to a single branch.

As explained with reference to FIG. 2A, the bias voltage can also be applied by means of an input and a pair of electrodes dedicated.
The modulator MEODS (Electro-Optic Modulator Double Output) of Figure 2C is dual-port RF type and dual optical output (dual drive ¨ dual output). It differs from the one presented by the figure 2A in the sense that the output Y coupler is replaced by a X directional coupler. The modulator MEODS thus comprises two optical outputs S01, S02, complementary to each other; phase shift between the two arms of the modulator changes the distribution of the irrensity between these two exits. The advantage of such a configuration is that it allows to use a balanced photodiode denoted PDE to convert the differential optical signal at the output of the modulator into an electrical signal and thus benefit from the benefits of such detection (increase of the amplitude of the RF output signal, suppression of the amplitude noise ...).
As shown in FIG. 2C, a photodiode Balanced comprises two optical inputs, consisting of fiber primers, respectively feeding two photodiodes connected in series with a RF output at the midpoint. For optimal operation of this type of photodiode, it is necessary to have a pairing in length of the fibers input which makes its implementation complex compared to a photodiode PIN (positive-intrinsic-negative) or UTC (from Uni-Traveling English) Carrier) standard to a single optical input. A pair of receivers Balanced photodiodes add to the output the signal strength differentials present on its two optical inputs and suppresses the signals common mode. In other words, it delivers a current corresponding to the difference of the currents generated respectively by the two photodiodes.
In the ideal case where the two photodiodes are perfectly identical (same sensitivity and equal fiber lengths), this causes a cancellation the continuous component (and associated noise). When using a standard optical-electrical converter (with single optical input), dominant noises are relative intensity noise (RIN) relative Intensity Noise) and the noise of shot (shot noise in English). The noise of shot is proportional to the average current detected, while the noise RIN is proportional to the square of this current. Like the signal-on-noise of an optical link increases with the square of the detected current, the increase in optical power is limited by the RIN that becomes so the predominant source of noise. A balanced detector allows you to delete the RIN and thus helps to improve the signal-to-noise ratio of the link.
Figure 2D illustrates another type of electronic modulator MEOIQ optical suitable for the implementation of the invention. This is a dual-control modulator called I / Q generally used for links with complex modulation formats such as modulation of amplitude in quadrature (in English, quadrature amplitude modulation:

QAM), QPSK phase modulations (in English, Quaternary Phase-Shift Keying) or m-PSK (m-ary Phase-Shift Keying), or modulation of Amplitude and Phase-Shift Keying or Asymmetric Phase-Shift Keying: APSK). It includes 2 modulators of Mach-Independent Zehnder integrated in a Mach-Zehnder interferometer primary. In the example of the figure, the polarization voltage of the primary Mach-Zehnder interferometer is applied to a third DC control input, which allows to control the relative phase between two arms of the interferometer, and therefore the output signal.
Figure 3 illustrates a frequency converter according to a first embodiment of the invention.
The first RF signal VRF1, having a central frequency f .RF1, and the second RF signal VRF2, having a center frequency fRF2, are, amplified and filtered by respective amplification and filtering blocks AF1, (optional) and then applied to the two RF inputs of an electronic modulator.
RF dual-port intensity optics (see, for example, FIGS. 2A, 2C, 2D). In FIG. 3 (and also in FIGS. 4, 5 and 6) the modulator ¨
possibly equipped with an RF combiner (CRF) - is designed so general by the MEO reference. The circuit and the bias voltage do not are not represented for the sake of simplification.
Two optical sources SL1, SL2, emitting at lengths 2v1, 22 are modulated by single frequency signals OL, SOL1, SOL2, of frequency foL1, mad respectively, generated respectively by local oscillators, OL1, 0L2. SO1 and SO2 the optical signals at wavelength X.1 and 12 respectively, of which the envelope is modulated at the frequency foLi, mad respectively.
The number of OL optical signals, equal to two, is given only as an example and is not limiting.
The sources SL1 may for example be semiconductor lasers.
directly modulated by the signals 0L1, 0L2 if the frequency of these is not too high (<15 GHz) or laser sources emitting in continuous mode, their optical flux being modulated by a modulator integrated or external controlled by signals SOL1, SOL2. The optical source and the local oscillator can also be integrated into an oscillator optoelectronics. The generation of OL1 and 0L2 signals in the form photonics can for example be made from a pulsed laser (or mode-locked laser mode) or a bi-laser the mode of which, respectively the period of the pulses or the difference Frequency between the two modes correspond to the LO frequency of interest.
The OL optical signals thus modulated are multiplexed by a multiplexer MUX, (ie WDM of the English Wavelength Division Multiplexing, that is wavelength division multiplexing) and injected at the optical input of the MEO modulator via a fiber optical system or, in the case of an integrated solution, a planar waveguide.
At the output of the electro-optical modulator two optical signals at the wavelengths Xi and X2 whose envelope has a complex modulation. The signal at wavelength 20 presents in particular spectral components at frequencies fRF1, fRF2, fL01. Similarly, signal at the wavelength X2 has particular components at frequencies fRF1, fRF2, fLO2 These two optical signals are then transmitted by optical fiber (or planar waveguide) to a DEMUX demultiplexer which allows them to be filtered in frequency and to direct them to photodetectors (or optical-electrical converters) COE1, COE2, typically PIN-type photodiodes (positive-intrinsic-negative) or UTC (from the English Uni-Traveling Carrier).
The electrical signals generated by photodetection of wavelengths 2%, 1, X2 contain the spectral components at frequencies fRF1 + f1, fRF1-f1, fRF2 + fLo1, fRF2-f1, 2fLoi, fRF1 + 2fLoi, and fRF11-FL02, fRF1-fL02, fRF2 + fL02, fRF2-fL02, 2fL02, fRF1 + 2fLo2, respectively. These spectral components of intermodulation appear at the time of the detection, because the photodetectors are sensitive to imensity bright, that is to say the square of the electric field.

They are then filtered using RF technology to extract the useful components and so reject others. For example, the converter of FIG. 3 makes it possible to convert the signal VRF1 to the frequency fRF1-fInd and the VRF2 signal at the frequency fRF2-fi_02. A single electro-optical modulator allows therefore to advantageously replace two RF mixers.
The set consisting of DEMUX demultiplexer and optical-electric converters COE1, COE2 form what we can call an optical-electrical conversion system rated SCOE.
In the embodiment of Figure 4, more than two Optical signals are injected into the MEO modulator. More precisely, in the converter of FIG. 4, four optical sources SL1, SL2, SL3, SL4 emitting at respective wavelengths Xi, X, 2, 13, X4 are modulated by single-frequency RF signals SOL1, SOL2, SOL3, SOL4 from crazy frequency, crazy, foL3, fired, generated respectively by oscillators premises, OL1, 0L2, 0L3, 0L4. The number of four is given only to as an example and is not limiting. This allows you to perform conversions multiple frequencies of the two RF signals VRF1 and VRF2. For example, the The converter of FIG. 3 makes it possible to convert the signal VRF1 to the frequency fRF1-fLoy but also at the frequency fRF1-ft_03, and the signal VRF2 at the frequency fRF2-fLO2 and at the frequency fRF2-fL04.
In addition, the embodiment of FIG. 4 differs from that of FIG. 3 in that the converter of FIG.
SDOE optical-electrical detection system different from that of the figure 3. This system includes a single DOE photodetector that generates a signal electrical complex whose useful spectral components are extracted by a bank of BFE electric filters. In this case, the wavelengths multiplexed must have sufficient frequency spacing important so that the heterodyne beat at the frequency difference between they do not generate a parasitic line in the useful band. This approach facilitates the use of a balanced photodiode for optical conversion electrical condition, provided that the electro-optical modulator is dual exit optical (see Figure 2C).

Of course, the optical-electrical conversion system of Figure 3 can also be used in a converter setting artwork more than two optical signals, and conversely that of Figure 4 can also be used in a converter implementing only two optical signals.
Figure 5 illustrates a frequency converter according to a third embodiment of the invention, implementing a modulator single RF access electro-optical and single optical output, such as illustrated in Figure 2B. The two RF signals VRF1 and VRF2, are combined by 0 the CRF radio frequency combiner before being applied to the single RF input of the modulator. The structure of the optical-electrical conversion system SCOE is identical to that shown in Figure 3 above and is consisting of a demultiplexer associated with as many photodiodes as there is of local oscillators.
Figure 6 illustrates a frequency converter according to a fourth embodiment of the invention, wherein a modulation at single sideband (SSB) is realized. For this, the signal VRF1 is divided into two components of the same amplitude by an RF separator marked SEP. One of these components is applied to the first RF input of the MEO modulator; the other component is out of phase, for example 90 (n / 2 radians) for a quadrature bias, and combined with the second radio frequency signal VRF2 by means of a DEP combiner / phase shifter before being applied to the second RF input of the MEO modulator (more generally, the value of the phase shift to apply to obtain an SSB modulation depends on the polarization voltage of the modulator). The presence of purely RF components such as SEP, DEP
limit the broadband nature of the converter, as it is also the case for the embodiment of Figure 5. The use of a modulation at single spectral band reduces the number of intermodulation products present in the photodetection signal.
Figures 7A-70 illustrate the performance of a frequency converter according to one embodiment of the invention, based on on the use of an electro-optical dual-access RF modulator (figure 2A) and in particular the importance of the choice of the bias voltage, and therefore the operating point of the modulator. The case considered corresponds to two RF signals, VFR1 and VRF2, with frequencies fRF1 = 28.5 GHz (band Ka) and fRF2 = 49.7 GHz (V band), and having the same amplitude equal to -10 dBm. For the sake of simplicity, a single +14.5 optical power signal dBm modulated at a foLi frequency of 10.2 GHz was used. The signal of output (envelope of the optical signal at the output of the modulator) is observed on a spectral range between 17 GHz and 22 GHz (useful band in Ka:
17.2 GHz ¨ 20.2 GHz), thus including the intermediate frequency fFi- = fRF, ¨
crazy.
Figure 7A shows the power spectrum of the RF signal from output, expressed in dBm, in the case where the bias voltage VB takes its optimal value Nha. This bias voltage value is optimal particularly for the first embodiment (described in FIG. 2A) in that it maximizes the rejection of the parasitic line at the frequency fFF2-fRF1, likely to be in the useful band. However, this tension of Optimal polarization can take other values, depending on the type of modulator used. It can be noted, in addition to the component useful to the ffi frequency a significant parasite component at a frequency 2foL1 and other lower parasitic components corresponding to products intermodulation at frequencies 2fRF1 + foL1-fRF2 and fRF2-3foL1. The component at frequency 2fou is not a problem because it is mono-frequency and can so be easily eliminated. Other intermodulation products, by against, may give rise to interference, so it is important that their level remains sufficiently low compared to that of the useful component.
Figures 7B and 7C correspond to situations in which the bias voltage deviates from 10 mV and 50 mV, respectively, of its optimum value. We can notice that this gap led to the emergence of a new, predominant intermodulation product at a frequency fRF2 - fRF1.
In FIG. 8, the square-shaped dots represent the relative power of intermodulation product at frequency f .RF2 - fRF1 by to that of the wanted signal at the frequency fFi (in dBc) as a function of the polarization voltage VB (the optimum value Vid2 being equal to 2.5 V), for the first embodiment. We can notice that this relative power increases rapidly with the difference IVB - Vit / 21, hence the importance of stabilizing VB. For VB = Vd2 the intensity of the component fRF2 ¨ fRF1 does not fall to zero, because the path is limited by the noise floor.
Figure 9 shows the power spectrum of the RF signal from output for the case VB = VII2 in the case of a modulator with double optical output followed by a balanced photodiode detection (Figure 20). It can be seen that the noise is reduced very considerably, to the point of not being visible in the figure, and that the parasitic component at the frequency 2f ff A
. RF1 +. YES -.RF2 faded away. These best performances, however, are obtained at the price a more delicate implementation especially due to the pairing in length fibers between the modulator and the optical receiver. .
For example, in future systems of Very High Speed Telecommunications (VHTS) Throughput Systems) the spectrum of gateway uplinks (gateways in English) may simultaneously use bands of frequencies Ka and V, while the downlinks (beams users) will be in Ka band. Each uplink will require 1 (Ka band) and 2 (V band) different frequency conversions. For example, for a repeater with 50 inputs from a gateway, the technology Conventional RF requires the implementation of 50 low-power amplifiers Low Noise Amplifier (LNA) in V-Band 50 demultiplexers, 100 converters from V-band to Ka-band and 50 Ka to Ka converters.
Figure 11 illustrates a SAT telecommunications satellite comprising a set of antennas receiving REC to receive uplink signals LM of a GW gateway, an antenna assembly in transmission TRAN to emit user beams forming a LD downlink and a CUT telecommunications payload realizing, among other things, frequency conversions and interconnections required.
Figure 10 is a block diagram of such a load useful, which can handle this situation using only 50 photonic frequency converters (vs.150 for RF technology) according to the invention. This has a positive and significant impact on the budgets of mass, power and cost of the satellite.
The payload of FIG. 10 is a forward path receiving the signals from 25 transmitting stations and rebroadcast them to the 150 beams users. The number of transmitting stations and beams is given to title example only. It includes a network of 25 antennas Al ¨ A25 operating, in reception, in the V bands (around 50 GHz) and Ka (around 30 GHz), on two orthogonal polarizations. So every antenna has two outputs, one for polarization and frequency band. Each Antenna output is filtered on the appropriate bandwidth, in V-band or in Ka band. Then rings of redundancy AR1, AR2 based on RF switches allow to implement additional channels in cold redundancy.
The assembly designated by the ECF reference comprises 50 photonic frequency converters according to the invention. Each electro-optic modulator this set receives on a first input RF a signal in band Ka, on a second between RF a signal in band V
and on its optical input three optical signals multiplexed at three lengths of distinct waves, 11, 12, 2L3 modulated at respective RF frequencies by local oscillator holes L01, L02, L03. At the output of each modulator, a photodetector allows to recover RF signals corresponded to different mixtures of RF frequencies received and generated by the local oscillators. The three useful signals are:
- the signal received in Ka band, now translated at 20 GHz about ;
the two received signals corresponding to the sub-bands of the V band, also translated at around 20 GHz.

At the output, rings of redundancy AR3, AR4, AR5 to base of RF switches allow to select the active channels, and RF filters filter the desired useful subband before to route the signals to the amplification section at 20 GHz AMP.
After power amplification, a last filter stage cleans up the spectrum before the signals are re-emitted by the antennas transmission to the 150 user beams.
FIG. 12 represents a photonic converter realizing simultaneously downconverting an RF signal, designated by the reference RFi, and the upconversion of a frequency signal intermediate Fi.
In the converter of FIG. 12, an electronic modulator MEO dual RF access optic is simultaneously powered on one of its RF channels by the RF1 signal and on the other RF channel by the Fi signal from an analog or digital PAN processor.
Two local oscillators L01, L02, of different frequencies, are transferred on optical carriers respectively to lengths of wave Xi and 12, which are multiplexed in wavelength and sent on the optical input of the modulator.
The first local oscillator L01, at the wavelength allows the downlink frequency conversion of the RF1 signal to a signal at a frequency Wire compatible with the input of the processor PAN. After demultiplexing (DEMUX) in wavelength, at the output of modulator, only the composite signal at the wavelength It carrying the component wire is isolated on one of the corresponding output paths of the demultiplexer. This signal is then detected by a converter optical / electrical, filtered (RF) and transmitted to the processor for treatment.
The second local oscillator L02, at wavelength 212 makes it possible to convert the rising frequency of the output signal Fi the PAN processor, at an intermediate frequency, to a radio frequency RF2 transmitted, after filtering, to the output amplification section of the payload (not shown). After wavelength demultiplexing, modulator output, only the composite signal at wavelength 12 carrying this RF2 component is isolated on one of the escape routes corresponding of the demultiplexer. This signal is then detected directly by an optical / electrical converter or can be transmitted by optical fiber to the RF filters and output amplification section of the payload.
In in this case, the optical / electrical conversion is performed upstream of the RF filters.

In the converter of FIG. 12, the electron modulator MEO optics is polarized in such a way that the amplitudes of parasitic components at frequencies RF1 Fi1 are significantly rejected with respect to the desired useful signal amplitude.

Claims (11)

21 A photonic frequency converter comprising:
an electro-optical intensity modulator (MEO) having an optical input (EO) and at least one optical output (SO) ,;
a set of optical sources (SL1 SL4) configured to generate optical signals (SO1 ¨ SO4) to at least two different wavelengths ¨ 24, modulated by oscillator signals (SOL1 ¨ SOL4) respective of which at least two have different frequencies;
an optical multiplexer (MUX) arranged for multiplexing said optical signals and injecting them into the optical input of the modulator;
- an optical / electrical detection system (SCOE, COE1, COE2, PDE) configured to convert optical signals from the electro-optic modulator of intensity towards the radiofrequency domain; and an electric filter bank (BFE) configured to extract spectral components of signals converted to the domain radio frequency;
characterized in that the electro-optical intensity modulator also has at least one input for a DC voltage and at least two RF inputs (EC1, EC2) to receive two signals modulation radiofrequency (VRF1 VRF2) at different frequencies. 2. Frequency converter according to claim 1 in which the so-called electro-optical intensity modulator (MEO) is of the Mach-Zehnder interferometer. Frequency converter according to claim 2 in which said electro-optical intensity modulator is of the type to interferometer dual-access Mach-Zehnder (MEODA) and two RF inputs (EC1, EC2) for two modulation radiofrequency signals (VRF1, VRF2) respectively. 4. Frequency converter according to claim 2 in which said electro-optical intensity modulator is of the interferometer Mach-Zehnder I / Q (MEOIQ) and has two RF inputs (EC1, EC2) for two respective modulation radiofrequency signals (VRF1, VRF2). Frequency converter according to claim 2 in which said electro-optical intensity modulator is of the interferometer Mach-Zehnder with Single RF Access (MEOAU) with Single Input RF (EC), the converter also comprising a signal combiner RF (CRF) configured to receive two RF modulation signals as input (V RF1, V RF2) respectively, combine them and apply them to said RF input of the electro-optical modulator of intensity. 6. Frequency converter according to one of Claims 2 to 5 also comprising a polarization circuit configured to apply to at least one modulator control input electro-optical intensity of a bias voltage (V B1, V B2, VB) suitable to minimize the intensity of a spectral component of the optical signals from the electro-optical intensity modulator at a sum frequency or difference in the frequencies of the modulation radiofrequency signals. 7. Frequency converter according to one of preceding claims also comprising a demultiplexer Optical device (DEMUX) arranged to demultiplex the optical signals from the electro-optical modulator of intensity. 8. Frequency converter according to one of preceding claims wherein said electro-optical modulator intensity is of the complementary dual optical output type (MEODS), the optical / electrical conversion system comprising at least one balanced photodetector (PDE). 9. Telecommunication payload (CUT) for a satellite (SAT) comprising at least one frequency converter according to one of the preceding claims. 10. Frequency conversion method by means of a converter according to one of the preceding claims comprising the following steps :
- apply to the RF input or inputs (EC, EC1, EC2) of the Electro-optical intensity modulator (MEO) of the two-signal converter radio frequency (V RF1, V RF2) whose frequencies must be converted;
- activate the optical sources of the converter (SL1 ¨ SL4) and for generating said optical signals (SO1 ¨ SO4) at least two different wavelengths (.lambda.1 ¨ .lambda.4), modulated by said signals local oscillators (SOL1 ¨ SOL4);
- recover optical signals from the optical output of the electro-optical intensity modulator, convert them to the domain radiofrequency and filter them to extract spectral components each corresponding to an input radio frequency signal mixed with one of said local oscillator signals. 11. The method of claim 10 further comprising the next step :
- apply to at least one control input of the electro-optical modulator of intensity a bias voltage (V B1, V
B2, VB) adapted to minimize the intensity of a spectral component of RF signals from the optical / electrical single-frequency detection system sum or difference of the frequencies of the radiofrequency signals of modulation.